Mendelian randomisation: a tool for assessing causality in observational epidemiology.

Area of Science:

• Epidemiology
• Genetics
• Public Health

Background:

• Establishing causality between risk factors and diseases is crucial for public health interventions.
• Observational studies often face confounding, making causal inference difficult.
• Unobservable confounding factors hinder accurate assessment of risk factor effects.

Purpose of the Study:

• To introduce and illustrate the Mendelian randomization method for inferring causality.
• To explain the assumptions and limitations of using instrumental variables in etiological epidemiology.
• To provide a framework for assessing the causal effect of modifiable risk factors on diseases.

Main Methods:

• Utilizing instrumental variables, specifically genetic variants (Mendelian randomization).
• Addressing confounding by leveraging the random assortment of genetic variants.
• Applying parametric assumptions for obtaining point estimates of causal effects.

Main Results:

• Mendelian randomization provides a robust approach to infer causality when confounding is present.
• Testing for causal effects is generally straightforward with this method.
• Obtaining precise effect estimates requires additional parametric assumptions.

Conclusions:

• Mendelian randomization is a powerful tool for etiological epidemiology and public health policy.
• The method helps overcome limitations of traditional observational studies in establishing causal links.
• Understanding the assumptions and limitations is key to its valid application.